00001 import roslib
00002 roslib.load_manifest('semantic_model_web_interface')
00003
00004 import mongo_ros as mr
00005 import semanticmodel.msg as smm
00006 import sensor_msgs.msg as sm
00007 import cv
00008 import cv_bridge
00009 import os
00010 from os import path
00011 from . import site_root, lock, DB_PORT, DB_NAME, DB_HOST
00012 import logging
00013 from logging import config
00014 import pymongo as pm
00015 import semanticmodel.db as sdb
00016
00017 config.fileConfig(path.join(site_root, 'logging.conf'))
00018 log = logging.getLogger('semantic_model_web.db')
00019
00020 cluster_collections = {}
00021 image_collections = {}
00022 bridge = cv_bridge.CvBridge()
00023
00024 POSITIONS = {
00025 'green_room': (20, 20, 8),
00026 'pool_room': (17, 35, 7),
00027 'white_lab': (34, 43, 10),
00028 'cafe': (30, 15, 7),
00029 'research_hallway': (5, 22, 7)
00030 }
00031
00032 def filter_query(q):
00033
00034
00035 return q
00036
00037 def get_cluster_collection(run):
00038 with lock:
00039 return get_cached_collection(cluster_collections, run, 'blobs',
00040 smm.BlobMessage)
00041
00042 def get_image_collection(run):
00043 with lock:
00044 return get_cached_collection(image_collections,
00045 run, 'images', sm.Image)
00046
00047 def get_cached_collection(colls, run, c, t):
00048 if run not in colls:
00049 colls[run] =\
00050 mr.MessageCollection(DB_NAME, sdb.collection_name(run, c), t)
00051 return colls[run]
00052
00053 def list_runs():
00054 conn = pm.Connection(host=DB_HOST, port=DB_PORT)
00055 db = conn[DB_NAME]
00056 runs = db['runs']
00057 return [('run{0}'.format(run['id']),
00058 '{0}.bag'.format(run['bag_file'])) for run in runs.find({})]
00059
00060 def run_info(run):
00061 conn = pm.Connection(host=DB_HOST, port=DB_PORT)
00062 db = conn[DB_NAME]
00063 runs = db['runs']
00064 id = int(run[3:])
00065 cursor = runs.find({'id': id})
00066 return next(cursor, None)
00067
00068 def image_file_loc(i, run):
00069 """
00070 @param i: Id of cluster
00071 @param run: Run id
00072
00073 Ensure that the file for the image of the cluster is available (download
00074 if necessary), and return its site-relative URL
00075 """
00076 with lock:
00077 coll = get_image_collection(run)
00078 meta = coll.find_one({'cluster_id': i}, True, sort_by='creation_time',
00079 ascending=False)
00080 relative_path = path.join('static/images',
00081 run,
00082 '{0}.jpg'.format(meta['blob_id']))
00083 filename = path.join(site_root, relative_path)
00084 log.debug("Looking for image file {0}".format(relative_path))
00085 if not path.exists(filename):
00086 log.info("{0} not found: getting from db".format(relative_path))
00087 msg, _ = coll.find_one(meta)
00088 img = bridge.imgmsg_to_cv(msg)
00089 d = path.dirname(filename)
00090 if not path.exists(d):
00091 os.makedirs(d)
00092 cv.SaveImage(filename, img)
00093 else:
00094 log.debug(" File found; not downloading")
00095 return '/'+relative_path
00096
00097 def next_id(coll, id):
00098 cursor = coll.find(filter_query({'id': {'$gt': id}})).sort('id', pm.ASCENDING)
00099 item = next(cursor, None)
00100 if item:
00101 return item['id']
00102
00103 def prev_id(coll, id):
00104 cursor = coll.find(filter_query({'id': {'$lt': id}})).sort('id', pm.DESCENDING)
00105 item = next(cursor, None)
00106 if item:
00107 return item['id']
00108
00109 def semantic_query(args):
00110 q = {}
00111 c = args.get('color')
00112 if c:
00113 q['color'] = c
00114 s = args.get('size')
00115 if s == 'small':
00116 q['diameter'] = {'$lt': 0.2}
00117 elif s == 'medium':
00118 q['diameter'] = {'$gte': 0.2, '$lt': 0.8}
00119 elif s == 'large':
00120 q['diameter'] = {'$gte': 0.8}
00121 shape = args.get('shape')
00122 if shape == 'planar':
00123 q['c01'] = {'$gt': 0.5}
00124 elif shape == 'curved':
00125 q['c01'] = {'$lte': 0.5}
00126 s = args.get('loc')
00127 if s:
00128 pos = POSITIONS.get(s)
00129 if pos:
00130 radius = pos[2]
00131 q['x'] = {'$lt': pos[0]+radius, '$gt': pos[0]-radius}
00132 q['y'] = {'$lt': pos[1]+radius, '$gt': pos[1]-radius}
00133
00134 return q
00135
00136
00137
